The Agent Skills Directory

[DATA_EXPOSURE]: The skill is instructed to explore the local codebase and documentation (READMEs, ADRs, architecture docs) to assess technical feasibility. This involves reading local files which is the intended primary purpose of the skill and does not include any network exfiltration logic.
[COMMAND_EXECUTION]: The skill saves a generated scope document to a specific local directory (./plans/<name>-scope.md). This is a localized file system write operation consistent with its documented purpose.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from the local codebase which could potentially contain malicious instructions.
Ingestion points: Local codebase files, architecture documentation, ADRs, and READMEs are read during the initial exploration phase.
Boundary markers: No explicit delimiters or instructions to ignore embedded instructions are used when processing the files.
Capability inventory: The skill is capable of reading local files, writing a markdown file to the ./plans/ directory, and interacting with the user via a structured question tool.
Sanitization: No explicit sanitization or validation of the ingested file content is described.

plan-feature